The Context#
AI, and especially LLMs like ChatGPT, are just about everywhere nowadays. Companies are integrating LLMs into things where many people don’t need or even want it. All those LLMs also need to be fed training data.
Public Open Source repositories like those on Github are a prime candidate.
How We Think About This#
While the OpenCommit Foundation acknowledges the many benefits brought by AI in general and in software development in particular, we also see the many negative sides and its pitfalls.
As such, we have many concerns ranging from enviromental impacts to copyright acknowledgements (or lack thereof) and moral obligations of AI companies towards open source and users.
Many concerns could be mitigated or even completely removed by appropriate behaviour from the side of AI companies and developers. We believe many of the problems associated with AI stem from poor behaviour. However, we do not expect to see any change in behaviour any time soon, especially from AI companies.
Due to the many concerns we have with the unrelenting behaviour of AI companies, we will be taking a hands-off approach to AI.
What We Commit To#
We commit to:
Actively combatting AI scraping
The platform uses various techniques to try and reduce, or preferably eliminate, the chances of your data being scraped by AI companies. Measures range from legal (Terms Of Use) to using tooling like Anubis.
We will not:
Add AI features to the platform
In the future we might add AI features but we will not add them if there is no clear use case for our users. If and when we ever introduce AI based features, they’ll be selectively and purposefully chosen.
Use your data for AI
We will never use your data, be it code repositories, issues or otherwise, for AI. Nor will we sell, trade or otherwise actively make any data available to AI.
Your data is yours.
Block AI generated code
Though we will not add AI features to the platform, we will not prohibit projects from making use of AI generated code. Whether or not a project accepts AI generated code is entirely up to the project.
What We Ask From You#
Quite simply we ask you, our users, to act responsibly with regards to AI. While we will not block anyone from using AI tools or AI generated content, we do ask that you consider the following points:
Clearly state your project’s policies around AI
For example, do you accept AI generated code? If so, under what conditions? We suggest that you mention this in your project README.
Indicate in your main README whether your project uses AI
We’d like to request project owners to make a note of their AI use, if any, in your project’s README.
Consider the impact of your AI use
For example: what is the environmental impact of your AI use? Could you have written the code yourself easily? Was the AI trained with or without proper permissions?
How This Might Evolve#
Like many technologies, AI changes rapidly. Though difficult to predict, we might end up introducing some AI technology to the platform in the future. If and when we do, we expect the negatives of that particular AI solution to have been resolved somehow.
What will definitely not change is: we will never sell, trade or otherwise use your data.
Your data is yours.